Application of change-point analysis to HPV infection and cervical cancer incidence in Xinjiang, China in 2011–2019

Abstract

Objective

Cervical cancer (CC), serving as a primary public health challenge, significantly threatens women’s health. However, in terms of change-points, there is still a lack of epidemiological studies on the incidence of HPV infection and CC in Xinjiang,China. This research aims to identify significant changes in the trends of HPV infection and CC prevalence in Xinjiang through change-point analysis (CPA) to provide scientific guidance to health authorities.

Methods

HPV infection and CC time-series data (from January 2011 to December 2019) were collected and analyzed. Meanwhile, their change-points were detected with binary segmentation method and the PELT method. Furthermore, patients were assigned into three groups based on their different ages and subsequently subjected to an analysis employing a segmented regression model (SRM).

Results

It was evident that for the monthly HPV time series, the binary segmentation method detected three change points in August 2015, February 2016, and September 2017 (with the most HPV cases). In contrast, the PELT method detected two change-points in September 2015 and April 2017 (with the most HPV cases). For the monthly CC time series, the binary segmentation method identified two change points in October 2012 and August 2019 (with the most CC cases), whereas the PELT method identified three change points in October 2012, August 2019 (with the most CC cases), and October 2019. The SRM demonstrated varying numbers of change points in distinct groups, with HPV infection and CC having the higher growth rate in the 30–49 and 40–59 age groups, respectively. Based on above results, this research was conductive to comprehending the epidemiology of HPV infection and CC in Xinjiang. In addition, it offered scientific guidance for future prevention and management measures for both HPV infection and CC.

1 Introduction

Cervical cancer (CC) emerges as a very frequent cancer in female individuals, characterized by high incidence and mortality [1]. According to estimates by the International Agency for Research on Cancer (IARC) [2], 660,000 new cases of CC was diagnosed worldwide and 350,000 of them died due to CC in 2022. The primary etiological cause of CC is recognized as persistent infection with high-risk human papillomavirus (hrHPV) [3, 4], which is primarily infected through various sexual contacts (like vaginal, anal, or oral sex). HPV infections are prevalent, affecting the majority of sexually active women throughout their lifetime [5]. However, HPV of more than 90% of individuals with HPV infections can be cleared spontaneously within a few years, and only persistent hrHPV infections can lead to invasive CC [6]. The progression from HPV infection to invasive CC can take up to 10 years, indicating that the CC occurrence is stemmed from multiple factors. Besides HPV infection, they may be associated with early sexual activity, multiple sexual partners, immunosuppression, smoking, oral contraceptive use, and poor socioeconomic conditions [7], among others. In addition, low-and middle-income countries (LMIC) face a significantly greater burden of CC than high-income countries [8, 9].

China, the world’s largest developing country, was reported to have an estimated 150,700 new cases and 55,700 deaths in 2022 [10], ranking fifth and sixth in female cancer incidence and deaths worldwide, respectively. With the increasing incidence and mortality, CC remains a major public health challenge that greatly affects women’s health in China. Furthermore, it has been reported that CC tends to be found in younger women [11, 12].

Xinjiang, situated in the northwest of China, is a vast and remote region comprising various ethnic groups, including Han, Uyghur, Kazah, among others. Limited awareness of HPV testing and prevention, along with inadequate hygiene conditions as well as a lack of systematic screening and prevention measures are closely linked to high HPV infection rate and CC among women in this area [13, 14].

Based on the time series data of HPV infection and CC cases in Xinjiang from 2008 to 2019, it is evident that HPV infection in this area is relatively severe, as depicted in Figs. 1 and 2 Number of individuals with HPV infections increased exponentially from 2008 to 2011 and then remained at a relatively high level. However, the number of CC cases exhibited an upward trend from 2008 to 2013 and then remained at a stable high level. Figures 3–6 display the HPV and CC cases for seven age groups in Xinjiang from 2008 to 2019. Data in these four figures reveal noticeable variation in the age distribution of HPV infection and CC incidence. It indicates that the women aged 30–49 possess the highest risk of HPV infection in Xinjiang, while those aged 40–59 years have the peak risk of CC. Moreover, the HPV and CC cases exhibited an overall downward trend in each group.

Figure 1

Number of HPV infection cases in Xinjiang in 2008–2019

Figure 2

Number of cervical cancer cases in Xinjiang in 2008–2019

Figure 3

Annual HPV infection cases of all ages in Xinjiang, 2008–2019

Figure 4

HPV infection cases in 7 age groups reported annually in Xinjiang, 2008–2019

Figure 5

Annual CC cases of all ages in Xinjiang, 2008–2019

Figure 6

Cases of CC in 7 age groups in Xinjiang, 2008–2019

Furthermore, the incidence of HPV infection and CC also varies among different ethnic groups [15, 16]. As illustrated in Figs. 7 and 8, a notable rise was observed in the percentage of CC cases among Uyghur women compared to the prevalence of HPV infections.

Figure 7

The proportion of HPV infection cases among various ethnic groups in Xinjiang, 2008–2019

Figure 8

The proportion of CC cases among various ethnic groups in Xinjiang, 2008–2019

In addition, a previous study [15] indicated that Uyghur women are at a higher risk of developing cervical cancer (CC) and dying from CC compared to women of other ethnic groups. This difference might be attributed to variations in lifestyle, social characteristics, and behavioral patterns [11].

Consequently, it is important to explore the distribution of HPV infection and CC in Xinjiang among different age groups and ethnic groups. Moreover, it is imperative to investigate the relationship and pattern of the above factors in HPV infection and CC prevention, intervention, and management. Hence, undertaking epidemiological studies in this area is crucial to facilitate developing more effective strategies to combat HPV infection and CC.

Change-point analysis (CPA) exhibits significant role in identifying the occurrence of a change [17, 18]. Generally, the time-series data may undergo obvious structural variations, which are also known as change points or breakpoints, segment the data. Currently, CPA has found extensive application in multiple scientific fields, such as finance [19], medical science [20], hydrology [21], bioinformatics [22], climate science [23–25], and more. In the field of public health, CPA also holds significant importance. Specifically, it can assist in identifying crucial change points during disease transmission or other health events, thereby contributing to formulation of enhanced intervention strategies and policies [26].

However, this method is ineffective in determining the change-points of COVID-19 [27], influenza-like illness [28], and malaria [29]. For example, Anwar [18] utilized the CPA to identify significant change points in outbreaks of lumpy skin disease in Africa, Europe, and Asia. Gargoum [30] employing a CPA, investigated the relationship between trends in human mobility and the number of individuals with COVID-19 infections as well as related mortality rates in distinct countries implementing various containment measures. The results make contribution to the control of the COVID-19 pandemic. However, there is limited research available on studying the significant changes in trends of HPV infection and CC prevalence using time-series CPA methods.

Identifying the change-points and trends in time-series data for HPV infection and CC is imperative to further investigate the prevalence of CC in the Xinjiang region, monitor its trends, and provide scientific guidance to public health authorities for the timely implementation of control measures. In this context, this research was developed to (i) discuss the epidemiological characteristics of HPV and CC incidence based on the specific change-points, and (ii) identify the change-points and variations in full HPV and CC time-series data as well as those in these data for women in varying age stages.

In conclusion, analyzing the change-points and trends in time-series data of HPV infection and CC can reveal the epidemiological characteristics and peak periods of CC, showing high significance. This provides a scientific basis for developing targeted prevention and control strategies. Additionally, this research explored the differences in HPV infection and CC incidence among women in different age groups. Consequently, this research is of significant importance for implementing targeted interventions and screening programs. By studying the change points and trends in HPV infection and CC, this research provides scientific evidence for the prevention and control of CC in Xinjiang and facilitates the development and implementation of public health policies.

2 Materials and methods

2.1 Study area and data source

The inadequate medical infrastructure in Xinjiang has resulted in insufficient sample collection efforts, leading to a scarcity of incidence rate data prior to 2010. Therefore, this research utilized the data on monthly cases suffering from HPV infection and CC in Xinjiang from January 1, 2011 to December 31, 2019. The statistical data from various affiliated hospitals of Xinjiang Medical University, Urumqi Maternal and Child Health Hospital, Xinjiang Uyghur Municipal People’s Hospital, The First Affiliated Hospital of Shihezi University, Hotan Prefecture People’s Hospital, Aksu Prefecture People’s Hospital, Tacheng City People’s Hospital, Yili Kazak Autonomous Prefecture maternal and child health care hospital, Kashgar Prefecture Second People’s Hospital, and Turpan Municipal People’s Hospital. The dataset covers the age, ethnicity, place of residence, diagnosis results, diagnosis time, and other related details of the patients. Since the data used for this analysis does not involve any private patient information, ethical approval or informed consent is not required.

2.2 The method of CPA

2.2.1 Detection of multiple change-points

CPA is used to identify the points that undergo statistical changes within a dataset. There are various algorithms to estimate the number and location of change-points, such as binary segmentation [31, 32], pruned exact linear time (PELT) [33], and segmented neighborhood [34, 35]. However, The change-point package [36] offers access to all these algorithms, allowing users to select the analysis methods that best fit their research needs.

In this research, the cpt.meanvar from the change-point package was employed, which was recognized for its capability to detect numerous change points. This function can detect the changes in mean and variance across four distinct data distributions: exponential, gamma, Poisson, and normal. By dividing data into m segments, the change-points can be identified by minimizing the ensuing function [36].

The two primary change-point algorithms used in this research were discussed in this section. The first refers to the Binary segmentation, which is a classic method for detecting change-points (Edwards and Cavalli–Sforza, 1965 [31];Scott and Knott in 1974 [32]). This method involves iteratively partitioning the dataset into two subsets and evaluating the cost on each side of the partition to determine the location of change points. Its simplicity and intuitiveness enable its extensive application in practice. The second refers to the PELT algorithm, which is proposed by Killick et al. in 2012 [33]. It is advantageous, because it not only can provide precise segmentation, but also can maintain a balance between accuracy and efficiency. It restricts the set of previous change-points, evaluates the cost of each segmentation by introducing a cost function, and seeks the change-points that can minimize the total cost [27].

Herein, $y_{1:n}=(y_{1},\ldots,y_{n})$ was assumed as a time series. Several change points (with the number m) were arranged for our model, and their positions were set as $t_{1:m}=(t_{1},\ldots,t_{m})$. Position of each change-point served as an integer between 1 and $n-1$ inclusive. Furthermore, $t_{0}=0$ and $t_{m+1}=n$ were defined. The change points were assumed to be ordered as $t_{i}< t_{j}$ and only if $i< j$. Consequently, based on these change points, the data were split into segments, and the ith segment contained $y(t_{i-1}+1):t_{i}$.

In consideration of m data segments, the change points can be identified to minimize the following function [36]:

$$ \sum_{i=1}^{m+1}\bigl[\mathcal{C} \bigl(y(t_{i-1}+1):t_{i}\bigr)+\beta f(m)\bigr]. $$

(1)

In the above formula, $\mathcal{C}$ represents a cost function for a segment and $\beta f(m)$ denotes a penalty to guard against overfitting.

2.2.2 Segmented regression model (SRM)[37]

Typically, the segmented relation can be represented by a covariate range that can be split in consecutive intervals. Within each interval or segment, the covariate exhibits a linear effect on response and remains continuous at all change points. This implies a persistent segmented linear behavior for the covariate-response relation. If $Y_{t}$ was selected to represent the cumulative number of cases at time $t=1,2,\ldots,n$, including linear term in the model can explain the relationship between the explanatory variable t and the mean response $E(Y_{t})$.

$$ E\bigl(\log (Y_{t})\bigr)=\beta _{0}+\beta _{1}t +\sum_{i=1}^{k}\gamma _{i}(t- \delta _{i})+. $$

(2)

Furthermore, this research assumes that there are $k+1$ regimes with the slope of $\beta _{1},\beta _{2}=\beta _{1}+\delta _{i}$, and the growth rate of $\beta _{k}=\beta _{1}+\sum_{1}^{k}\gamma _{i}$ from the slope can be calculated, i.e., $\gamma _{k}=\exp \{\beta _{k}\}-1$ for each regime $k=1,2,\ldots,k+1$. In addition to the growth rates, the period it took to double the number of cases is a very simple parameter which can be easily understood and often adopted by epidemiologists and health experts in their reports. These doubling times (DTs) can be acquired through simple parameterization of the slopes. Thus, $d_{k}=\log (2)/\beta _{k}$ can be given for each regime. The estimates of all model parameters, including breakpoints, are obtained by maximizing Poisson likelihood or quasi-likelihood. When fitting multiple piecewise models, the Bayesian Information Criterion (BIC) can be adopted to determine the model that is most applicable for the data. It means that the model should be equipped with an appropriate number of breakpoints to fit the observed data. The SRM can be fitted based on the “segmented package”.

3 Results

3.1 Data processing and analysis or description

The data employed herein were filed using EXCEL 2022 and subjected to CPA process through the change point (cpt.meanvar) and segmented packages in R software. Furthermore, the significance level was set at 0.05.

3.2 Change point in the time-series data of HPV and CC

In this research, the monthly HPV and CC time-series from 2011 to 2019 were analyzed employing both binary segmentation method and PELT method for the purpose of change-point detection. The two methods exhibited little difference in the number of detected change points, but quite similar positions of these change points.

For HPV time-series, the binary segmentation method identified three change-points, leading to four corresponding segments, as demonstrated in Fig. 9(A). Each segment represents the average number of HPV cases during the corresponding period (red line). These three change points occurred in August 2015, February 2016, and September 2017. Additionally, the PELT method detected two change-points in September 2015 and April 2017, corresponding to three segments in the time-series, as illustrated in Fig. 9(B).

Figure 9

Change points in monthly time-series of HPV in Xinjiang, 2011–2019, detected by the binary segmentation method (A) and PELT method (B)

The time-series data related to CC exhibited two change-points after detection using the binary segmentation method, which facilitated the recognition of three corresponding segments Fig. 10(A). Each segment represents the average number of CC cases during the corresponding period (red line). These two change-points occurred in October 2012 and August 2019, respectively. Moreover, the detection employing the PELT method revealed three change-points, thereby enabling the identification of four corresponding segments within the time-series Fig. 10(B). Specifically, the change-points were identified in October 2012, August 2019, and October 2019, respectively. The positions of all change-points detected by two methods are summarized in Table 1.

Figure 10

Change points in monthly time-series of CC in Xinjiang, 2011–2019, detected by the binary segmentation method (A) and PELT method (B)

Table 1 The position of change points detected by two methods

By combining with our research findings, it is observed that with the gradual and continuous progress of HPV testing and CC screening programs in Xinjiang, the detection rates have been effectively improved. This improvement better reflects the trends of HPV infection and CC in Xinjiang.

3.3 The results of SRM

Due to the low incidence of HPV infection and CC among women under 20 years old, and considering the differences in incidence rates of HPV infection and CC among women of various ages, the sequence of HPV cumulative counts was categorized into three age groups (20–29, 30–49, and 50+). Meanwhile, sequence of CC cumulative counts was also classified into three age groups (20–39, 40–59, and 60+). Moreover, the line segment fitting was employed to examine the number of cases between two change-points within intervals. Above classifications and categories facilitate the analysis of parameters within a certain timeframe, such as change points, growth rates, and doubling times.

Regarding the HPV data, the optimal model was selected on the basis of the lowest BIC in Table 2, leading to variations in the count of change points across different age groups. Specifically, in Table 2, the age group of 20–29 exhibits the lowest BIC value at the third change point, while the age groups of 30–49 and 50+ demonstrate the lowest BIC values at the fifth and third change points, respectively.

Table 2 BIC values of CPA in each sequence (HPV) for women of different ages

Figure 11 displays the CPA results based on Table 3, serving as the points connected by lines, which are deemed as segments with similar trends, marked in varying colors. For each segment, the incidence and corresponding 95% confidence intervals (CI) were computed, as presented in Table 3 below.

Figure 11

The result of CPA for sequence of HPV infection cumulative outcomes counts in three age groups (20–29, 30–49, and 50+): (A) 20–29; (B) 30–49; (C) 50+

Table 3 CPA incidence in each sequence (HPV)

As observed in Fig. 11(A), three change-points was detected in the 20–29 age group, yielding four segments. The incidence rate for each segment is as follows: 1.9% (95% CI: 1.5% ∼ 2.4%), 4.5% (95% CI: 4.3% ∼ 4.6%), 2.9% (95% CI: 2.7% ∼ 3.2%) and 0.2% (95% CI: 0.1% ∼ 0.3%), respectively. In contrast, five change points were identified in the 30-49 age group, as illustrated in Fig. 11(B). The incidence is 10.8% (95% CI: 10% ∼ 11.7%) in segment 1, which was relatively high. In addition, the incidences for other segments are as follows: 5.9% (95% CI: 5.6% ∼ 6.2%), 1.7% (95% CI: 1.6% ∼ 1.7%), 3.9% (95% CI: 3.8% ∼ 3.9%), 2.3% (95% CI: 2.1% ∼ 2.4%), and 0.2% (95% CI: 0.1% ∼ 0.3%), in that order. Furthermore, there are three change points in the 50+ age group, as explicated in Fig. 11(C). The corresponding incidences are as follows: 1.6% (95% CI: 1.4% ∼ 1.7%), 5.3% (95% CI: 5.1% ∼ 5.4%), 2.4% (95% CI: 2.2% ∼ 2.6%), and 0.3% (95% CI: 0.2% ∼ 0.4%), respectively, in sequence. Additionally, the lowest incidence is found in the last segment for all age groups.

Using the same principle, the CC data in Table 4 reveal that the age group of 20–39 exhibits the lowest BIC value at the second change point, while the age groups of 40–59 and 60+ have the lowest BIC values at the fifth and third change points, respectively.

Table 4 BIC values of CPA in each sequence of women at different ages in Xinjiang (CC)

The CPA results in Fig. 12 correspond to the data in Table 3, which are presented as points connected by lines. Herein, the incidences and corresponding 95% CI value for each segment are presented in Table 5 below. As given in Fig. 12(A), in the age group of 20-39, the incidence is 3.7% (95% CI: 3% ∼ 4.3%) in segment 1, 2.3% (95% CI: 2.1% ∼ 2.5%) in segment 2 and 1.1% (95% CI: 1% ∼ 1.1%) in segment 3. There exist five change points in the age group of 40–59, as explicated in Fig. 12(B). The incidence is high in segment 1, reaching 9.4% (95% CI: 7.5% ∼ 11.4%). While that for other segments is as follows: 6.1% (95% CI: 5.6% ∼ 6.6%), 3.2% (95% CI: 2.9% ∼ 3.4%), 2.1% (95% CI: 1.9% ∼ 2.3%), 1.6% (95% CI: 1.4% ∼ 1.7%)and 1.1% (95% CI: 1% ∼ 1.2%) in segments 2, 3, 4, 5, and 6, respectively. Three change points were identified in the age group of 60+, as detailed in Fig. 12(C). The corresponding incidence for each segment is as follows: 5.4% (95% CI: 4.4% ∼ 6.4%), 3.1% (95% CI: 2.6% ∼ 3.6%), 2% (95% CI: 1.9% ∼ 2.1%), and 1.3% (95% CI: 1.2% ∼ 1.4%) in sequence, respectively. Overall, for all age groups, the datasets exhibit the same decreasing trend.

Figure 12

The result of CPA for the sequence of CC cumulative outcomes counts in three age groups of 20–39 (A), 40–59 (B), and 60+ (C)

Table 5 CPA incidence in three age groups of women in Xinjiang (CC)

Doubling times (DT) is utilized in various fields, including epidemiology, to describe the time it takes for a quantity to double under a specified set of conditions. The Shorter DT signifies an elevated disease incidence. Tables 6 and 7 report the DT with the 95% CI for each segment. For the time-series of monthly HPV data outlined in Table 6, the age group of 30-49 shows a high incidence, the DTs for different segments are 6.73,12.06, 42.21, 18.31, 30.91, and 361.88, respectively. For the time-series of monthly CC data, as presented in Table 7, DTs are observed to be short in the first segment in all age groups and gradually become shorter over time, signifying a high incidence. Notably, for the age group of 40–59, the DTs for various segments are as follows: 7.71, 11.68, 22.17, 33.42, 44.22, and 63.38, respectively. However, the DT for each group in both sets of data increases, which is certainly a positive sign from an epidemiological perspective.

Table 6 DT and the 95% CI of each segment for three age groups (HPV)

Table 7 DT and the 95% CI of each segment for three age groups (CC)

DT for HPV infections is shorter in the age group of 30–49, while that for CC is shorter in the age group of 40–59, indicating higher growth rates in these groups. The higher growth rate of HPV infection among women aged 30–49 may be attributed to their sexually active lifestyle, which greatly increases the opportunities for persistent HPV infections. The growth of CC incidence in middle-aged and elderly individuals may be stemmed from decreased immune function, declined ovarian function, and significant hormonal fluctuations with age. This weakens their ability to eliminate and suppress the virus, resulting in activation of latent viruses or recurrent HPV infections, among other factors.

In Tables 8 and 9, the segmented trends across various age groups were summarized using the average percentage change. It was computed by averaging the slopes, weighted by the corresponding interval width. Herein, the growth rate was averaged as. In Table 8, the HPV dataset for women in the age group of 30–49 reflects a higher growth rate, reaching 3.16% (95% CI: 3.11% ∼ 3.19%), those in the age groups of 20–29 and 50+ signify the growth rates of 2.50% (95% CI: 2.45% ∼ 2.56%) and 2.42% (95% CI: 2.38% ∼ 2.45%), respectively. As presented in Table 9, the CC dataset for women in the age group of 40–59 also reflects a higher growth rate at 2.91% (95% CI: 2.83% ∼ 2.98%), while those for the age groups of 60+ and 20-39 indicate the growth rates of 2.32% (95% CI: 2.25% ∼ 2.39%) and 1.95% (95% CI: 1.87% ∼ 2.03%), respectively ($AGE=e^{Est.}-1$).

Table 8 The means of the average percent change (est.) and the AGE of women with various ages in Xinjiang (HPV)

Table 9 The means of the average percent change (est.) and the AGE of women with various ages in Xinjiang (CC)

4 Conclusion

The World Health Organization (WHO) advocated for the elimination of CC globally in 2018 and released the “Global Strategy to Accelerate the Elimination of CC” in 2020 [38], proposing to achieve the medium-term goal of “90-70-90” by 2030. This goal involves several key measures, such as HPV vaccination, CC screening, and intervention. As a country with a significant burden of CC globally, China has witnessed varying degrees of increase in CC incidence and mortality in recent years, with a trend of onset occurring at younger age. The distribution pattern indicates that women in rural areas experience higher incidence and mortality than urban areas [39]. Therefore, women in rural areas, especially those in impoverished regions with limited healthcare resources, should be focused in terms of prevention and control efforts for CC in China. Xinjiang, with its vast land area, substantial rural population, lower economic level, and relatively underdeveloped medical conditions, currently exhibits higher levels of HPV infection and CC incidence compared to the national average [40]. It becomes evident that analyzing the temporal changes in HPV infection and CC incidence in Xinjiang over time can provide valuable epidemiological information, exhibiting high regional significance.

Based on the CPA algorithm, key change points (three) for HPV infection in Xinjiang were identified with the same binary segmentation method. They occurred in August 2015, February 2016, and September 2017, and exhibited an upward trend each year. Two significant change points in the trend of CC prevalence were identified in October 2012 and August 2019, persisting at a high level for a long time. China has extensively implemented the National Cervical Cancer and Breast Cancer Screening Program for Rural Women nationwide (referred to as the “Two Cancers Program”) since 2009. Thanks to this program, Xinjiang has initiated HPV testing projects in 11 counties/cities in 9 conditional regions since 2014, and started to offer CC screenings free of charge for rural women in 85 counties/cities across 13 regions during 2015-2016 [41]. The effect of implementing these strategies shows that with the gradual and continuous progress of HPV testing and CC screening programs, their detection rates have been effectively elevated, better reflecting the trends of HPV infection and CC in Xinjiang. However, the global infectious disease COVID-19 at the end of 2019 brought great inconveniences for daily lives of people, such as disruptions to medical check-ups and treatments. In this case, relevant health authorities had to shift their focus to respond to the pandemic, resulting in a decline in CC detection rates. Based on the results obtained from SRM in this research, it is evident that the DT for HPV infection in women aged 30–49 years is shorter, indicating a higher growth rate of HPV infection. Similarly, DT for CC of women in the 40–59 age group is also shorter, signifying a higher growth rate of CC within this age group. Firstly, women aged 30–49 are typically in their sexually active period, significantly increasing the risk of recurrent HPV infections. Secondly, women in this age range experience a higher probability of being detected through gynecological examinations and screenings. It is a gradual process from persistent HPV infection to CC that may span several decades, making CC more likely to occur post-menopause, especially for women aged 45–60. Aligning with our research findings, CC demonstrates a higher growth rate in women aged 40–59 years. This may be associated with distinct factors, such as age-related decline in immune function, difficulties in clearing HPV infections, declined ovarian function, and notable hormonal fluctuations leading to the activation of latent viruses or recurrent HPV infections. Furthermore, with the exacerbation of population aging trend, the proportion of elderly women affected by CC continues to rise [14]. Analyzing the change points and trends can reveal the disease patterns, thus aiding in formulating more effective and scientific prevention and control strategies. In this research, the key Change-points in data associated with HPV infection and CC incidence for women in Xinjiang were identified using statistical methods. Furthermore, importance of prevention and control measures tailored to various age groups was highlighted by analyzing the disparities in growth rate across different age groups. The results unveil that HPV infection and CC exhibit high incidences in Xinjiang, which may be caused by joint effect of multiple factors, such as geographical location, economic status, cultural diversity, and ethnic composition. Therefore, it is crucial to prioritize the dissemination of knowledge related to CC in the region, foster healthy sexual behaviors, prevent HPV infections, as well as ensure early detection, diagnosis, and treatment to mitigate the burden of CC. CC is preventable, and effective implementation of tertiary prevention exerts a critical effect in reducing the incidence and mortality associated with CC. However, owing to the large population in China, preventive HPV vaccination is hard to be included in national immunization program, leading to a low vaccination rate among women. According to the estimates of cumulative HPV vaccine coverage rates among women aged 9–45 in 2020 [42], the reported vaccination rates in Beijing, Shanghai, and Zhejiang were 8.28%, 7.37%, and 4.68%, respectively. In contrast, Tibet, Qinghai, and Xinjiang reported the vaccination rates of 0.06%, 0.39%, and 0.46%, respectively. The above data suggest that Xinjiang exhibits a higher incidence of HPV infection and CC, but relatively low HPV vaccination rates. As a result, strengthening tertiary prevention becomes especially crucial in this context. Herein, we give the following recommendations for comprehensive prevention persistent HPV infection and CC. The primary preventive measures encompass health education (promoting condom use and personal hygiene, along with limiting sexual partners), encouragement of receipt of HPV vaccination to prevent HPV infections and related CC. The secondary preventive strategies comprise promoting and standardizing the CC screening to guarantee early identification, diagnosis, and intervention of lesions. In addition, the tertiary preventive actions are composed of active treatment of high-grade lesions.

In conclusion, preventing and controlling HPV infections and CC requires adopting integrated specific approaches tailored to women with varying ages. This necessitates the joint effort of multiple sections, such as healthcare departments, medical professionals, and the public. Executing the suggestions given above can efficiently alleviate the CC-induced burden in Xinjiang and promote the health conditions of women there. Future prevention and control efforts can be developed in terms of formulating the change-point identification methods with higher accuracy to investigate the epidemiological features of HPV infection and CC incidence. Additionally, assessing the efficiency of preventive and controlling measures through analysis remains essential.